TABLE 4.

Summary of extracted features from accelerometric and acoustic signals for aspiration risks classification.

Modality	Domain	Feature	Reference
Accelerometry	Time	Statistical features (mean, median, variance, skewness, kurtosis, Interquartile Range), Absolute difference between mean and median Stationary, normality, dispersion ratio Significance level of normality, Maximum hyolaryngeal excursion, proportion of signal corresponding to maximum hyolaryngeal excursion Zero-crossing Energy, entropy rate, Lempel-Ziv complexity Linear prediction coefficient	Lee et al. (2006)
			Lee et al. (2011)
			Merey et al. (2012)
			Shu et al. (2022)
	Frequency	Peak frequency, centroid frequency Band width Peak Fast Fourier Transform magnitude, frequency at spectral peak Frequency corresponding to max spectral density over time of the short-time Fourier transform Difference between frequency corresponding to 75% and 25% of max spectral density at time corresponding to max frequency Statistical features of power spectrum	Lee et al. (2011)
			Merey et al. (2012)
			Shu et al. (2022)
	Time-frequency	Wavelet entropy and energy Wavelet packet coefficient Relative energy and entropy for wavelet decomposition	Lee et al. (2011)
			Sejdic et al. (2013)
			Shu et al. (2022)
Acoustics	Time	Statistical features (standard deviation, skewness, kurtosis) Entropy rate, Lempel-Ziv complexity Phase-space thresholding, Normalized energy of the 3rd quartile of average power Jitter and shimmer features	Sarraf Shirazi et al. (2012)
			Sarraf Shirazi et al. (2014)
			Park et al. (2022)
			Shu et al. (2022)
	Frequency	Peak frequency, centroid frequency Band width	Frakking et al. (2022)
			Shu et al. (2022)
	Time-frequency	Wavelet Entropy Wavelet packet coefficient	Sejdic et al. (2013)
			Shu et al. (2022)